Benchmark Almanac

ALBA

Gemma 3 12B 51.1 leads since February 2026
Not in model cards

Models answer 800 European Portuguese prompts across eight areas, including wordplay, grammar and sound. A language model grades the answers.

The AMALIA model report compares 14 models. It rescales judge ratings from 1 to 5 onto a 0 to 100 scale.

Released
February 2026
Kind
Knowledge in many languages
Level
School
Scored
Graded by a model
Answers
Open-ended
Tasks
All public
Measures
Capability
Language
Portuguese
In and out
Text → Text
Interaction
Single turn
Items from
Written by experts
Refresh
One fixed set
Run by
Anyone
Strongest model tested
Qwen3-8B #154 on the capability index
Last model added
September 2025
Scores refreshed
November 20th, 2026
14
Entries in its paper, 8 models we track

Gemma 3 12B leads with 51.1 among models released by September 2025 Gemma 3 12B leads with 51.1 among models released by September 2025

Best score per entry from the benchmark's paper, read on November 20th, 2026. A paper’s table doesn’t add newer models. Where a model ran with several settings or harnesses, we keep its best. Names in gray aren't in our model list: fine-tunes, agent systems, products and older models. They don't count toward leaders. This view shows only models we track, each at its best.

The newest model with a score here came out in September 2025, and nobody has tested a newer one. Among the 20 highest-ranked models on the capability index, these have no score yet:

  1. 1 Gemma 3 12B 51.1
  2. 2 Qwen3-8B 49.8
  3. 3 AMALIA-9B-DPO 43.6
  4. 4 Gemma 2 9B 41.1
  5. 5 Apertus 8B Instruct 38.7
  6. 6 EuroLLM-9B 38.5
  7. 7 AMALIA-9B-SFT 37.1
  8. 8 Ministral 8B 35.6
  9. 9 Llama 3.1-8B 31.3
  10. 10 Gervasio-8B 31.1
  11. 11 Qwen2.5-7B 31.0
  12. 12 Salamandra-7B 27.4
  13. 13 Mistral 7B v0.1 21.7
  14. 14 OLMo 2-7B 16.9

Source: The ALBA paper

#EntryScoreSettingDateSource
1 Gemma 3 12B 51.1 Mar 2025 Official leaderboard
2 Qwen3-8B 49.8 Apr 2025 Official leaderboard
3 AMALIA-9B-DPO open weights 43.6 Official leaderboard
4 Gemma 2 9B 41.1 Jun 2024 Official leaderboard
5 Apertus 8B Instruct 38.7 Sep 2025 Official leaderboard
6 EuroLLM-9B open weights 38.5 Official leaderboard
7 AMALIA-9B-SFT open weights 37.1 Official leaderboard
8 Ministral 8B 35.6 Oct 2024 Official leaderboard
9 Llama 3.1-8B 31.3 Jul 2024 Official leaderboard
10 Gervasio-8B open weights 31.1 Official leaderboard
11 Qwen2.5-7B 31.0 Sep 2024 Official leaderboard
12 Salamandra-7B open weights 27.4 Official leaderboard
13 Mistral 7B v0.1 21.7 Sep 2023 Official leaderboard
14 OLMo 2-7B open weights 16.9 Official leaderboard
#ModelScoreSettingDateSource
1 Gemma 3 12B 51.1 Mar 2025 Official leaderboard
2 Qwen3-8B 49.8 Apr 2025 Official leaderboard
3 Gemma 2 9B 41.1 Jun 2024 Official leaderboard
4 Apertus 8B Instruct 38.7 Sep 2025 Official leaderboard
5 Ministral 8B 35.6 Oct 2024 Official leaderboard
6 Llama 3.1-8B 31.3 Jul 2024 Official leaderboard
7 Qwen2.5-7B 31.0 Sep 2024 Official leaderboard
8 Mistral 7B v0.1 21.7 Sep 2023 Official leaderboard

Number one over time

Lab claims count from the day the card came out, independent runs from the model’s release.

Google DeepMind
Benchmark paper Not tested since Feb 2026

Source: The ALBA paper

Scores over time

All the scores we have, one dot each, so a model tested at several settings appears more than once.

Plot with other benchmarks
In the benchmark’s paper
20.0 30.0 40.0 50.0 60.0 2024 2025 20.0 30.0 40.0 50.0 60.0 2024 2025

Source: Official ALBA leaderboard

No lab model card reports it yet

We read the headline results of 540 launch posts from 57 labs, and none of them list this benchmark.

More in Multilingual

All 545 in Multilingual
  • MMMLU In use MMLU professionally translated into 14 languages. Multiple-choice accuracy. Claude Mythos Preview 92.7%
  • WMT24++ In use English-to-55-language translation of WMT24 test data. Scored with automatic metrics such as MetricX or xCOMET. K-EXAONE 90.5%
  • MMLU-ProX In use MMLU-Pro translated into up to 29 languages with expert review. Multiple-choice accuracy per language. Qwen3.7-Max 87%
  • CLIcK In use Models answer 1,995 multiple-choice questions on Korean language and culture. Questions come from official Korean exams and textbooks in eleven categories. Solar Open 2 90.7%
  • Global MMLU Lite 3.0 In use Models answer 400 human-translated MMLU questions per language in 23 languages. Version 3.0 adds Czech, Hungarian, Oriya, Slovak and Tajik and updates Italian. Accuracy. Claude Opus 5 95.7%
  • MMMB In use Models answer multiple-choice questions about images in six languages across 15 categories. The set has about 12,000 questions. LFM2.5-VL-3B 83%

Data audit

8 of 9 checks pass, 1 warning. Look first at the Newest score check. The newest came 14 months ago, in September 2025, for Apertus 8B Instruct. Every benchmark

Catalog record 2 pass
  • Pass Task access All tasks are public.
  • Pass Confidence Checked against a primary source.
Scores 1 warning 2 pass 2 notes
  • Pass Independent runs Scores from the benchmark’s paper (14 entries).
  • Pass Models scored 8 models we track, plus 6 other entries on its leaderboard.
  • Warning Newest score The newest came 14 months ago, in September 2025, for Apertus 8B Instruct.
  • Note Frontier models None of the top 10 has a score. The strongest tested is Qwen3-8B (#154).
  • Note Headroom Scored as a plain number, with no fixed ceiling.
Leaderboard import 3 pass
  • Pass Importer We read the board on aclanthology.org: 8 rows for models we track and 6 other entries at the last read.
  • Pass Last read November 20th, 2026.
  • Pass Read worked The last read worked.
Consistency 1 pass 1 note
  • Note Lab claims No model has both a lab number and an independent run.
  • Pass Known issues No published audit found problems.

How we check